Revealing Formation Mechanism of Self‐Induced InAlN Core‐Shell Nanorods by Aberration‐Corrected TEM

Abstract

The interest in nanoscale structures has gained immense momentum due to their scientific and technological potentials. Among accessible nanostructures, semiconductor materials fabricated as one‐dimensional nanorods (NRs) offer fascinating physical properties and engineering capabilities for active components in present and future nanoscale functional devices [1]. In particular, group‐III nitride semiconductor NRs based on AlN, GaN, InN and their ternary alloys are attractive due to the widely tunable direct bandgap (0.64‐6.2 eV), high crystal quality and improved light extraction efficiency. Recently, self‐induced core‐shell ternary nitride NRs have been demonstrated [2‐5]. However, the understanding of the formation mechanism and self‐induced separation remains debated. In this work we investigated the formation mechanism of self‐induced core‐shell InAlN NRs growth on an amorphous C substrate. A number of InAlN samples were grown by magnetron sputtering epitaxy directly onto TEM grids which supports amorphous C films (substrates) with varying growth time. Apart from the otherwise identical growth conditions, the growth time (t) was varied in the steps of t=1, 2, 3, 5, and 20 min for the different samples. The obtained samples were subsequently investigated in plan‐view projection using the doubly‐corrected Linköping FEI Titan 3 60‐300. The microscope is equipped with a monochromated X‐FEG high‐brightness gun, efficient high solid angle Super‐X EDX detector and ultrafast Gatan GIF Quantum ERS post‐column imaging filter. A time series of InAlN NR formation, at different nucleation and growth stage, imaged by plan‐view STEM‐HAADF together with corresponding SAED patterns, is shown in Fig. 1. The imaging conditions strongly promote image contrast dependence on the mass (Z number) and sample thickness with reduced diffraction contrast contributions. The bright features exhibit high mass and indicate that they are In‐enriched. Al and In elemental distribution in the grown samples were examined using EDX elemental mapping as shown in Fig. 2. The elemental maps corroborate that the bright features (in Fig. 1) are In‐enriched (InAlN) islands, while less bright areas represent Al‐rich (InAlN) islands in the 1 and 2 min samples. The situation is somewhat more complex for the samples with extended growth time (3, 5, and 20 min), where the thickness contribution to the STEM‐HAADF and EDX intensities must also be considered. Additionally, we performed EDX quantification, SAED pattern analysis, high‐resolution TEM and STEM‐HAADF imaging as well as statistical STEM‐HAADF image contrast analysis. By investigating different InAlN sample at different growth time we were able to follow NRs evolution process: from initial In‐rich InAlN seed nucleation to final NR core‐shell formation. This enabled us to derive NR formation scenario as a function of the growth time, see Fig. 3. The NR formation process can be divided into two distinct regimes. I) the nucleation and coalescence phase where In‐enriched islands are surrounded by Al‐rich, and II) the growth phase during which the In‐enriched islands develop into cores and the surrounding Al‐rich environment develop into shells around the cores, and the core‐shell structure eventually reach a steady state NR growth. To account for the present observations we consider a number of factors affecting the NRs formation, including: adatom (In, Al, and N) surface kinetics (adsorption, desorption and surface diffusion), chemical potential, surface energy, thermal stability, and incoming flux (shadowing effect) during dual magnetron sputter epitaxy.